I'm a founding product person — I do the product judgment and I build the thing, from "is this worth solving" to a shipped system with real eval numbers behind it.

The decision I chose to validate against Stockfish rather than heuristics — objective truth over plausible-looking output. Move quality is not something a rulebook should be guessing at when a real engine can just answer the question.

Proof 236 automated tests (as of 2026-07-24). The app itself validates every move against a real Stockfish binary — that's the architectural decision. The suite that checks the app is contract-tested against a fake engine implementing the same interface, with expected values (golden fixtures) seeded from genuine Stockfish runs, plus one test that spins up the real engine end-to-end.

The decision The gate's job is coarse sieving, not quality ranking — so the metric that matters is false-drop rate, not accuracy. A gate that quietly discards good signal fails invisibly, and accuracy alone would hide exactly that failure.

Proof Measured against a 30-example hand-labeled golden set (as of 2026-07-10): 76.7% accuracy, 73.3% precision, 78.6% recall — and a 20% false-drop rate against a <10% target. The target is not met, and the labeling surfaced a systematic bias I hadn't written into the prompt. A reusable eval harness is the next build, precisely so this gets re-measured rather than estimated.

The decision I picked a narrow buyer and built the distribution, not just the software — building was never the hard part.

Proof A shipped platform — deployed, tested, and running — plus a real pipeline behind it: a targeted buyer list built out in a live CRM with a written outreach playbook, so the software has somewhere to go.